Multi-Region Deployment for Global Web Application

Multi-Regional Deployment of a Web Application

Multi-regional deployment reduces latency for users in different geographic zones, ensures failover for regional outages, and can be required by data localization rules (GDPR, 152-FZ).

Architectural Patterns

Active-Passive — one region is primary, second gets traffic only on failure. Simple, but failover not instant (15–60 seconds).

Active-Active — both regions take traffic simultaneously. More complex: need data sync, write conflict resolution.

Read Replicas — writes only in primary region, reads from nearest. Suits most read-heavy web apps.

AWS Multi-Region with Route 53

# Terraform: two EKS clusters
module "eks_eu" {
  source  = "terraform-aws-modules/eks/aws"
  region  = "eu-west-1"

  cluster_name    = "myapp-eu"
  cluster_version = "1.29"
  vpc_id          = module.vpc_eu.vpc_id
  subnet_ids      = module.vpc_eu.private_subnets

  managed_node_groups = {
    main = {
      instance_types = ["m6i.xlarge"]
      min_size       = 2
      max_size       = 10
      desired_size   = 3
    }
  }
}

module "eks_us" {
  source  = "terraform-aws-modules/eks/aws"
  region  = "us-east-1"
  # similarly
}

# Route 53: Latency-based routing
resource "aws_route53_record" "api" {
  zone_id = var.hosted_zone_id
  name    = "api.mysite.com"
  type    = "A"

  set_identifier = "eu-west-1"
  latency_routing_policy {
    region = "eu-west-1"
  }

  alias {
    name                   = aws_lb.eu.dns_name
    zone_id                = aws_lb.eu.zone_id
    evaluate_target_health = true
  }
}

Database: Replication Between Regions

PostgreSQL with logical replication:

-- PRIMARY (eu-west-1)
CREATE PUBLICATION myapp_pub FOR ALL TABLES;

-- REPLICA (us-east-1)
CREATE SUBSCRIPTION myapp_sub
  CONNECTION 'host=eu-primary.rds.amazonaws.com user=replicator password=secret dbname=myapp'
  PUBLICATION myapp_pub;

AWS Aurora Global Database — managed variant:

Primary Region: eu-west-1 (write + read)
Secondary:      us-east-1 (read-only, ~1s lag)
Secondary:      ap-southeast-1 (read-only)

Aurora Global failover: ~1 minute, automatic via Route 53.

Kubernetes: Multi-Regional Deployment

# ArgoCD ApplicationSet: deploy to multiple clusters
apiVersion: argoproj.io/v1alpha1
kind: ApplicationSet
metadata:
  name: myapp
  namespace: argocd
spec:
  generators:
    - list:
        elements:
          - cluster: eks-eu-west-1
            region: eu-west-1
            db_host: aurora-eu.cluster.rds.amazonaws.com
          - cluster: eks-us-east-1
            region: us-east-1
            db_host: aurora-us.cluster.rds.amazonaws.com
  template:
    metadata:
      name: 'myapp-{{region}}'
    spec:
      project: default
      source:
        repoURL: https://github.com/myorg/myapp
        targetRevision: HEAD
        path: helm/myapp
        helm:
          values: |
            region: {{region}}
            database:
              host: {{db_host}}
      destination:
        server: '{{cluster}}'
        namespace: myapp

Stateless Application

For multi-regional deployment the app must be stateless:

// DON'T store state in process memory
// BAD:
const sessions = new Map<string, Session>(); // lost on restart

// GOOD: Redis (with replication)
import { Redis } from '@upstash/redis';

const redis = new Redis({
  url: process.env.UPSTASH_REDIS_URL!, // Upstash supports multi-region
  token: process.env.UPSTASH_REDIS_TOKEN!,
});

async function getSession(sessionId: string): Promise<Session | null> {
  return redis.get<Session>(`session:${sessionId}`);
}

Vercel/Netlify: Edge Network

For Next.js/Nuxt simplest multi-regional deploy — Vercel Edge Network:

// Server components and API routes deploy as Edge Functions
// automatically to 30+ regions worldwide

// app/api/config/route.ts
export const runtime = 'edge'; // deploy to edge nodes worldwide

export async function GET() {
  const region = process.env.VERCEL_REGION ?? 'unknown';
  return Response.json({ region });
}

Latency overhead for EU→US: 100–150ms. Edge deployment: 5–20ms to user.

Monitoring Multi-Regional Deployment

# Prometheus: metrics by region
labels:
  region: eu-west-1
  cluster: eks-eu

# Grafana dashboard: Request Rate by Region, Latency by Region
# Alertmanager: alert if one region unavailable

Setting up Active-Active multi-regional deployment in two AWS regions with Aurora Global Database — 5–10 working days.